Claude Platform debuts on AWS, giving customers console and API access
- Anthropic and AWS made Claude Platform on AWS generally available on May 11, letting customers use Claude’s native console and APIs through AWS accounts. - The key twist is architectural: Anthropic runs the service and processes data outside AWS’s security boundary, while billing, IAM, and CloudTrail stay in AWS. - That gives enterprises a Bedrock alternative with fuller Claude features, but it also sharpens the tradeoff between convenience and stricter residency controls.
Anthropic just changed how Claude shows up inside AWS. Not by replacing Amazon Bedrock, but by adding a second path that gives AWS customers direct access to Anthropic’s own platform — console, APIs, beta features, managed agents, code execution, the whole thing — without making them leave their AWS account. That matters because a lot of enterprise AI friction is not model quality. It’s procurement, identity, billing, and audit trails. This launch tries to remove that friction while keeping Anthropic’s first-party product intact. ### What actually launched? Claude Platform on AWS went generally available on May 11, 2026. AWS customers can now sign up through the AWS console and use Anthropic’s native platform experience through AWS-linked authentication and billing instead of setting up a separate direct Anthropic relationship. Anthropic’s own release notes frame it as the Claude API and platform becoming accessible through AWS on Anthropic-managed infrastructure. (aws.amazon.com) ### Is this just Bedrock with a new label? No — and that’s the whole point. In Amazon Bedrock, AWS operates the inference stack for Claude models. In Claude Platform on AWS, Anthropic operates the service, and customers get the same API surface and platform features they would get from Anthropic directly, with AWS acting more like the enterprise wrapper around access, billing, and governance. (aws.amazon.com) ### What do customers get that’s different? They get the first-party Anthropic experience without the usual account sprawl. The feature list includes the Messages API, Files API, Message Batches API, Claude Managed Agents, Agent Skills, code execution, tool use, and early-access beta features. The AWS marketplace listing also points to extras like web search, MCP connectors, prompt tools, evaluation tools, and model comparison inside the Claude Console. (platform.claude.com) ### So what stays inside AWS? Identity, billing, and audit plumbing. Customers use existing AWS IAM credentials and policies, usage rolls into AWS billing, spend can count against existing AWS commitments, and CloudTrail provides audit visibility. Basically, the control plane feels familiar to an AWS enterprise buyer even though the AI service itself is not being run the same way Bedrock runs it. ### What’s the catch? (platform.claude.com) The catch is data handling. AWS says Claude Platform on AWS is operated by Anthropic and that customer data is processed outside the AWS security boundary. Anthropic’s docs make the same distinction. So this is not the obvious choice for teams with strict regional residency or boundary requirements. AWS explicitly positions it for customers that want Anthropic’s native platform and do not have those specific constraints. (aws.amazon.com) ### Why would a company want this instead of direct Anthropic access? Because enterprises hate duplicate systems. If a company already runs purchasing, identity, tagging, permissions, and logging through AWS, this setup lets it adopt Claude without creating a separate vendor workflow for every team. There’s still an Anthropic organization under the hood, but signup through AWS provisions it in a way tied to the AWS account, and requests are scoped through AWS-linked workspaces. (aws.amazon.com) ### Why does this matter beyond Anthropic? Because it hints at where enterprise AI distribution is going. The old split was simple — use the model maker directly for the newest features, or use the cloud provider for governance and easier buying. Claude Platform on AWS blurs that line. Anthropic keeps the native product experience. AWS keeps the enterprise relationship. That’s a stronger bundle than either side offering only half the stack. (docs.aws.amazon.com) ### Bottom line? This is less a new model launch than a new sales and control-plane design. Anthropic is trying to meet enterprises where they already live — inside AWS — without giving up ownership of the product surface. If that works, expect more AI vendors to copy the pattern. (aws.amazon.com) (claude.com)
